Some patients with diabetes just don’t fit the mold: the thin young kickboxing enthusiast on a low-carb diet whose blood glucose is persistently elevated or the baby not yet 6 months old. Consider, too, the patient who appears to have type 1 or 2 diabetes, but he or she doesn’t respond well to the treatment plan.
Genetic testing may help determine whether these outliers are among the approximately 1% to 2% of patients with diabetes who have monogenic forms of the disease, including maturity-onset diabetes of the young (MODY), neonatal diabetes, and some types of syndromic diabetes. But most never get the necessary testing, said Toni Pollin, PhD, MS, a former genetic counselor and now associate professor of medicine at the University of Maryland School of Medicine.
“It’s atypical to be offered genetic testing,” Pollin explained.
In everyday clinical situations, genomic testing and analysis of the resulting data often aren’t well integrated into routine practice. But Pollin hopes the Personalized Diabetes Medicine Program (PDMP) pilot will help change that. The pilot is part of the National Institutes of Health–funded Implementing Genomics in Practice Network (IGNITE). The Network along with others like the Sync for Genes program, which is funded by the Office of the National Coordinator for Health Information Technology (ONC), aims to overcome the technical and practical challenges of integrating genomics into everyday clinical practice.
To help identify and funnel more patients with MODY and other forms of monogenic diabetes into appropriate care, Pollin and her colleagues have cast a wide net. They use self-screening questionnaires distributed in person or online by practices in Maryland and Pennsylvania, as well as online outreach and taking physician referrals.
Participants who are likely to have monogenic diabetes undergo sequencing of 40 different genes linked with these conditions. Variants identified as clinically significant are confirmed in a Clinical Laboratory Improvement Amendments or College of American Pathologists (CLIA/CAP)–accredited laboratory. The laboratory generates a report that is shared by a study physician and genetic counselor with the patient and his or her clinician and is integrated into the electronic medical record (EMR).
Patients are monitored after diagnosis to find out whether their treatment changes or their condition improves and how they experience these events. Pollin hopes the evidence collected will contribute to future practice guidelines on who and when to test for MODY. Such recommendations could go a long way toward sparing patients from harmful or ineffective treatment resulting from misdiagnoses.
For example, patients with MODY2, or glucokinase MODY, often have mild glucose elevations, are not at risk of the complications linked with other forms of the disease, and usually do not require treatment, Pollin explained. In fact, misdiagnosis with type 1 or type 2 diabetes and subsequent treatment with insulin may pose a bigger threat by triggering hypoglycemia and even low-birth-weight neonates.
The team also is building EMR-based clinical decision support tools that will alert clinicians when testing for monogenic forms of diabetes is indicated. For example, most cases of permanent neonatal diabetes in infants younger than 6 months are monogenic, and about half of these have genetic causes that respond markedly better to high-dose oral medications than to insulin injections, Pollin noted.
“A lot of people will go their entire careers without seeing a baby with diabetes,” she said. “It lends itself to an alert in the medical record so that if they do, the baby won’t have to rely on the clinician remembering a lecture to be assured access to genetic testing.”
A number of similar pilot programs are currently under way as part of the IGNITE Network, including one aimed at improving hypertension treatment for black patients.
Black patients with hypertension have a 2- to 3-fold higher risk of developing chronic kidney disease, something recent studies suggest is linked to variations in the APOL1 gene. An IGNITE pilot, which is led by researchers from Mount Sinai Hospital in New York City, will enroll black patients with hypertension from primary care clinics in Harlem and the Bronx to study ways to implement APOL1 testing and related decision support into care and determine how it affects outcomes.
Another IGNITE pilot led by Indiana University researchers will work to integrate pharmacogenomic testing for 24 widely used drugs into the primary care practices serving underserved populations.
To make the most of genomic information in practice, clinicians need convenient access to relevant genomic information about their patients.
Patients increasingly have part or all of their genomes sequenced either through direct-to-consumer services or during the course of clinical care, generating vast amounts of data. Among the 3 billion base pairs that make up the human genome, there are 59 clinically actionable genetic variants, Geoffrey Ginsburg, MD, PhD, director of the Duke Center for Applied Genomics and Precision Medicine, explained.
“You are obviously not going to put a whole genome in the electronic medical record,” said Ginsburg. “But you really want to have the metadata that is relevant to clinical decision making.”
Currently, these data often wind up in the EMR as a PDF or a fax that isn’t easy for clinicians to use, explained Gil Alterovitz, PhD, an assistant professor of biomedical informatics at Harvard Medical School. Alterovitz, who also is the lead investigator of the Sync for Genes program, and his colleagues are currently leveraging the same technology that drives apps like Facebook and Uber to deliver genomic information to clinicians in a more useful, standardized way.
“Sync for Genes is the first step towards integrating genomic information,” Alterovitz said.
The project will use Application Programming Interfaces (API) that EMR systems or clinical applications can query for relevant data. For example, when you search Facebook for cars for sale in your city, Facebook sends a query to its API requesting data on cars for sale in your specified area.
Sync for Genes will also use a standard language for genomic information that builds on the Fast Healthcare Interoperability Resources (FHIR) that already provide a consistent way for EMRs to communicate clinical information. Standards are essential for interoperability and organizing any type of electronically transmitted data, Alterovitz explained.
“It leverages workflows clinicians are already using to access genomic information,” Alterovitz said.
But presenting genomic data to physicians is not enough.
“A lot of physicians don’t have the background to understand what to do with genomic information, so clinical decision support capabilities need to be established,” said Ginsburg, who leads IGNITE Network’s coordinating center. These tools need to alert physicians when a patient has a genetic variant that affects their health, what the consequences of the variant are, and what can be done about it.
The Sync for Genes pilots will provide valuable test cases for some of the more technical aspects of genomic data sharing and use. For example, the US Food and Drug Administration is running one of these pilots that will explore sequencing quality and the use of genomics for regulatory purposes. “It helps us learn using a real-world application with real players,” said Alterovitz.
Integrating family history into EMRs is another way of leveraging genomic data in practice. Often, family history can alert physicians to the possibility of a genetic condition and the need for genetic testing. For example, in one case a man who indicated his mother and maternal grandmother died of breast cancer was tested and found to be a BRCA carrier, according to Lori Orlando, MD, of Duke University Medical Center, who is a principal investigator of IGNITE’s family history pilot along with Ginsburg. His children also were tested and 2 of the 3 were found to be a carrier, which has implications for their own cancer risk or their children’s.
Too often, however, EMRs don’t contain a thorough family history. A 2012 query of data from the Epic EMR at Medical College of Wisconsin, which included 721 000 patient encounters, found that 85% lacked a family health history and only 1% of records had 3 generations of data, said David Dimmock, MD, medical director of the Rady Children’s Institute for Genomic Medicine in San Diego. “A lot of physicians don’t have time to collect an adequate family history,” Ginsburg explained. “They don’t always do it well and even when they do it well they may not know how to interpret the results [and how they relate to clinical guidelines].”
So, Orlando and Ginsburg are testing an electronic family history tool called MeTree. The pilot has enrolled more than 2500 patients from 35 health clinics in 6 states. MeTree helps a patient fill out a detailed family history outside of an office visit. The program then creates an evidence-based report for the patient’s physician and patient that is integrated into the EMR. The pilot is currently assessing how these electronic histories affect patient screening, behavior, and outcomes.
“It’s trying to drive a very productive patient-physician interaction,” Ginsburg said.
Intermountain Healthcare in Utah also has launched a separate family history pilot as part of Sync for Genes. The pilot guides patients through an electronic pedigree within the health system’s patient portal, said Stan Huff, MD, the health system’s chief medical informatics officer.
Once a pedigree is filled out, the software taps into existing risk algorithms to directly advise patients on how they can reduce their risks. It also provides a report physicians can review with the patient.
“We are trying to be able to collect an accurate family history, do clinical decision support, and risk analysis to provide better guidance to patients,” Huff said.
So far, Intermountain’s tool has received mixed reviews from patients.
“Everybody who uses it is excited about it,” Huff said. “But it is not used by 100% of patients. They are not all interested in entering their information.”
In the long run, Huff and his colleagues would like to find a better way to integrate family history data into EMRs, and with patient and family members’ permission autopopulate some family health data. He explained that Intermountain Healthcare includes about 60% of health clinics in the state.
“There is a high likelihood we have [a patient’s] parents’ health records,” he explained. “A lot of information could be filled in by the computer. The challenge is security and privacy.”
One free electronic family history tool already is available from the US Surgeon General’s Office. Physicians who would like to learn more about family history, pharmacogenomics, or other ways to apply genomic medicine can also consult a toolbox created by the IGNITE Network.
“One could argue family history is one of the most powerful genetic tools we have today,” said Ginsburg.
Note: The print version excludes source references. Please go online to jama.com.
Kuehn BM. Pilot Programs Seek to Integrate Genomic Data Into Practice. JAMA. Published online July 12, 2017. doi:10.1001/jama.2017.7181